On June 23, 2016, an extreme tornado struck the Yancheng City of the Jiangsu Province, China. It was the most serious tornado to hit China in the past 50 years, with many schools, flats, and factory buildings severely destroyed. The tornado claimed 99 lives, injured approximately 800, and affected more than 1.6 million people. ^{Reference Lyu, Wang and Cheng1} Beyond the financial and physical damages, individuals who experienced natural disasters could suffer from severe and prolonged psychological consequences, such as posttraumatic disorder, depression, and anxiety. ^{Reference Neria, Nandi and Galea2–Reference Fan, Zhang and Yang4} Adolescents have been studied extensively by trauma researchers have been found to be a particularly vulnerable group. ^{Reference Alisic, Jongmans and van Wesel5–Reference Zhou, Wu and An8} Some adolescents reported positive psychological changes following stressful and traumatic experiences, known as posttraumatic growth (PTG). ^{Reference Zhou and Wu9,Reference Zhou, Wu and Zhen10}

PTG refers to positive changes that occur from a person’s struggles with highly challenging life crises, ^{Reference Calhoun and Tedeschi11} including positive changes in self, interpersonal relationships, and life attitude. ^{Reference Tedeschi and Calhoun12} A few studies have found that adolescents experiencing traumatic events are likely to exhibit PTG. For example, a study showed that adolescents affected by the Ya’an earthquake in China, reported a 76.2% prevalence of PTG. ^{Reference Chen and Wu13} Similarly, after 2008 Wenchuan earthquake, 60.2% of adolescents reported PTG. ^{Reference Zhou, Wu and Zhen14} In another study following the 2016 Yancheng tornado, some adolescent survivors also demonstrated PTG. ^{Reference An, Yuan and Zhang15} Considering the widespread and positive existence of PTG among posttraumatic adolescents, many researchers take considerable interests in exploring the factors affecting PTG, such as perceived social support, ^{Reference Yuan, Goh and Xu16} positive cognitive coping, ^{Reference An, Ding and Fu17} and deliberate rumination. ^{Reference Zhang, Xu and Yuan18} In this study, we expected to conduct an in-depth study about factors promoting PTG among adolescent survivors, empirically investigate the positive psychological responses following trauma, and promote adolescents’ mental health after traumatic events.

Parents are important sources for adolescents’ ability to cope with traumatic experiences, ^{Reference Man, Liu and Xue19,Reference Duru and Balkis20} and Tian et al. found that parental attachment significantly positively predicted PTG. ^{Reference Tian, Wu and Wang21} The quality of teenagers’ relationships with their parents can significantly influence their mental health outcomes, ^{Reference Helsen, Vollebergh and Meeus22–Reference Woodhouse, Ayers and Field24} especially when they experience adversities or traumatic events. ^{Reference Song, Thompson and Ferrer25} Parental attachment is defined as the affectional ties that binds adolescents to their parents. ^{Reference Bowlby26} Previous studies found that individuals possessing secure attachment reported less psychological problems following traumatic events. ^{Reference Fraley, Fazzari and Bonanno27–Reference Cohen and Katz29} Moreover, securely attached individuals are more likely to demonstrate PTG. ^{Reference Mikulincer, Shaver and Horesh30} These individuals may have more positive views and better interpersonal relationship, which was found to be helpful for them to handle traumatic events. ^{Reference Arikan, Stopa and Carnelley31} Yuan found that parental attachment predicted PTG in adolescents. ^{Reference Yuan, Xu and Liu32} However, to our knowledge, the mechanism through which parental attachment influences PTG is still not clear. The current study aims to fill in the gap.

Following trauma, individuals’ resilience deserves attention. Resilience refers to an individual’s ability to experience stress or traumatic events with positive adaptation instead of traumatic outcomes, ^{Reference Masten and Cicchetti33} which is positively associated with PTG. ^{Reference Bensimon34} The theory of resilience indicates resilient people are more likely to seek help from others, balance their emotions, and find new meanings of the self, others, and the world when dealing with traumatic events. ^{Reference Arikan, Stopa and Carnelley31,Reference Seligman and Csikszentmihalyi35} Besides, the resilience-promoting model reveals resilience can help people alleviate the severity of traumatic response. ^{Reference Masten and Monn36} Thus, resilience could promote individuals’ positive changes in self, relationships, and life attitude after adversity. There is evidence showing that adolescents with higher resilience were less influenced by the tornado and presented higher PTG. ^{Reference Yuan, Xu and Liu32}

In addition, early parental attachment may affect adolescents’ response to traumatic experiences. Previous studies have shown that secure attachment was positively associated with higher resilience. ^{Reference Mikulincer, Shaver and Horesh30,Reference Karreman and Vingerhoets37,Reference Shibue and Kasai38} Attachment theory argues that the quality of early attachment relationship determines individuals’ internal working models of self and others. ^{Reference Bowlby26} The working models then influence individuals’ relationships and coping with stressors. ^{Reference Bartholomew and Horowitz39} Securely attached people hold positive working models, and it may lead to better interpersonal relationships, ^{Reference Arikan, Stopa and Carnelley31} which could help individuals to adjust successfully and adapt positively when they experience adversity and traumatic events. ^{Reference Hobfoll, Dunahoo and Monnier40} These positive processes may eventually lead to PTG. In other words, positive parental attachment could lead to adolescents’ resilience in the context of significant adversity, ^{Reference Yuan, Goh and Xu16} which may in turn affect the development of PTG. Thus, we expect resilience to be a mediator that explains the relationship between parental attachment and PTG among adolescents.

There have been studies reporting the concurrent relationships among the parental attachment, resilience, and PTG in adolescents by using the cross-sectional data following the tornado. ^{Reference Yuan, Goh and Xu16,Reference Zhou, Zhen and Wu41} However, the longitudinal effects of parental attachment on resilience and PTG remain unknown. A longitudinal study is necessary to examine the causal links among the 3 variables. In this study, we recruited adolescent survivors who experienced the 2016 Yancheng tornado and assessed parental attachment at 12 months (T1), and resilience and PTG at 18 months (T2) after the tornado. Based on the theoretical models and empirical studies mentioned above, we proposed two hypotheses: (1) parental attachment at T1 could positively predict PTG at T2; (2) parental attachment at T1 could indirectly influence PTG at T2 by means of increased resilience at T2.

Results

Descriptive Statistics and Correlations Among Main Measures

The means, standard deviations, and correlations among the 3 variables are shown in Table 1. Parental attachment at T1 positively correlated with resilience at T2 (r = .331; P < 0.001), and posttraumatic growth (PTG) at T2 (r = .193; P < 0.001). Besides, resilience at T2 was positively associated with PTG at T2 (r = .384; P < 0.001).

Table 1. Descriptive statistics and correlation analysis of parental attachment at T1, resilience at T2, and PTG at T2

Abbreviations: PA-Alie, parental attachment-alienation; PA-Com, parental attachment-communication; PA-Tru, parental attachment-trust; PA, parental attachment; RES-Str, resilience-strength; RES-Opt, resilience-optimism; RES-Ten, resilience-tenacity; RES, resilience; PTG-PS, posttraumatic growth-personal strength; PTG-NP, posttraumatic growth-new possibilities; PTG-RO, posttraumatic growth-relating to others; PTG-SC, posttraumatic growth-spiritual changes; PTG-AL= posttraumatic growth-appreciation of life; PTG, posttraumatic growth.

*P < .05.

**P < .01.

***P < .001.

Mediating Effect of Resilience

At first, a measurement model was built, including 3 latent variables: parental attachment at T1, resilience at T2, and PTG at T2. The parental attachment latent variable was evaluated by alienation, communication, and trust toward parents. The resilience latent variable was evaluated by strength, optimism, and tenacity. The PTG latent variable was evaluated by spiritual changes, appreciation of life, relating to others, new possibilities, and personal strength. In the measurement model, we assumed that there was a correlation between parental attachment at T1 and PTG at T2, and the factor loadings of the latent variables were freely estimated. The model was found to fit the data well (χ²/df = 2.641; NFI = .955; IFI = .971; TLI = .961; CFI = .971, RMSEA [90% CI] = .068 [.053-.084]), indicating that the measurement model was acceptable and could be used for subsequent structural model analysis.

Then, to examine the mediating effect of resilience at T2 on the effect of parental attachment at T1 on PTG at T2, we built a direct effect model with the link from parental attachment to PTG. This model also fit the data well (χ²/df = 2.557; NFI = .972; IFI = .983; TLI = .974; CFI = .983; RMSEA [90% CI] = .067 [.044-.090]). Path analyses revealed that parental attachment was a significant positive predictor of PTG.

At last, we established an indirect effect model to examine the mediating role of resilience. Results showed that the model again fit the data well, (χ²/df = 2.641; NFI = .955; IFI = .971; TLI = .961; CFI = .971; RMSEA [90% CI] = .068 [.053-.084]). Figure 1 shows the path coefficients of the model. Analysis of each path in the model revealed that parental attachment at T1 positively predicted resilience at T2 (β = .38; P < 0.001) and resilience at T2 can predict PTG at T2 negatively (β = .36; P < 0.001), while the direct effect of parental attachment at T1 on PTG at T2 was no longer significant. Thus, resilience at T2 fully mediated the relationship between the parent attachment at T1 and PTG at T2.

Figure 1. The mediation model from parental attachment at T1(12 months post event) to PTG at T2(18 months post event) through resilience at T2 among adolescents following the 2016 Yancheng.

We further conducted bias-corrected bootstrap tests with 2000 replications to test the significance of the indirect effect. First, 2000 bootstrap samples were created from the original data. Then, an estimate of 2000 mediating effects were generated and saved; the effect values were sorted by size; the 95% median confidence interval is estimated using the 2.5^th percentile and the 97.5^th percentile. In this situation, if the 95% confidence interval of an indirect path coefficient does not include 0, the indirect path is indeed significant. Table 2 shows the bias-corrected bootstrap test results. After testing, the 95% confidence interval for the indirect effect from parental attachment at T1 to PTG at T2 through resilience at T2 did not include zero, indicating a significant mediating effect (Table 2).

Table 2. Bias-corrected bootstrap tests for the mediating effect

Abbreviations: PA, parental attachment; Res, resilience; PTG, posttraumatic growth.

***P < .001.